A prospective study of temperamental factors in adolescents' vulnerability to substance abuse will be conducted with a sample of 1,800 subjects, surveyed initially in 6th grade and followed at 1-year intervals from 7th grade through 10th grade. Temperament dimensions of activity, emotionality, and sociability using the Revised Dimensions of Temperament Survey (DOTS-R) as the primary measure of temperament and including measures of sociability and generalized self-control. Data from adolescent subjects will be obtained through questionnaires administered in classrooms. Indicators of substance abuse will (a) early onset, (b) rapid escalation, (c) high-intensity use, and (d) problematic consequences of use. Measures of aggression, impulsivity, and risk taking will be obtained and tested for combined effects with temperamental factors. The protective factor of parental support will be investigated as buffer of temperamental characteristics; socioeconomic status and family disruption will be investigate as factors that may exacerbate vulnerability from temperamental characteristics. Clustering analysis will be used to determine patterning of substance use over time and variables discriminating adolescents who remain at minimal experimentation levels from those who show steady escalation of substance use. Mediational models will be tested in cross-sectional and longitudinal analyses to determine whether temperament relates to vulnerability through affective, coping, or social mechanisms. The respective contributions of temperament and conduct disorder to substance use will be tested in cross-sectional and longitudinal analyses, and structural modeling analyses will investigate the nature of relationships between temperament and conduct disorder over time. The results will have implications for the theory of vulnerability to substance abuse and for the design and targeting of prevention programs.